Multirobot Persistent Patrolling in Communication-Restricted Environments

In multirobot patrolling, a team of robots is deployed in an environment with the aim of keeping under observation a set of locations of interest. In several realistic mission scenarios, only human operators sitting at a base station are able to assess the situation on the basis of data sent by robots. Examples include watching pictures or video streams to detect intruders and correlating measurements to detect leaks of contaminants. We assume that a communication infrastructure is available only in some regions of the environments, from where messages can be exchanged with a sufficient bandwidth between the robots and the base station. In this paper, we first extend the classical multirobot persistent patrolling framework and the related idleness evaluation metrics to such environments with a limited number of "communication zones". Then, we present some centralized and distributed patrolling strategies tailored for this communication-restricted framework. Finally, we evaluate their performance using ROS/Stage simulation.